The point of building out an eXtreme Scale deployment is to make more memory, CPU, and network bandwidth available. More memory means we can store more objects. More CPU cycles gives us lower latency in servicing client requests and running grid agents which will be introduced in Chapter 7. More network bandwidth allows more clients to simultaneously connect to the grid by spreading our data across more physical computers. We'll do a brief overview of the topics here, and get into more detail as we progress.
A distributed eXtreme Scale deployment is made up of at least one catalog server process and at least one container server process. The catalog server coordinates the efforts of the container processes. A common scenario is to have a few catalog servers managing many container servers in a deployment.
In a distributed deployment, a container process is where our objects live after a client commits a transaction with put or insert operations. Catalog and container processes...